Biochemistry 1980, 19, 43 13-4321
4313
Characterization of the Enzyme Complex Involving the Folate-Requiring Enzymes of de Novo Purine Biosynthesist Gary K. Smith,* W. Thomas Mueller,s Gail Folena Wasserman, William D. Taylor, and Stephen J. Benkovic*
ABSTRACT: Evidence is presented for a functional association of G A R TFase and the trifunctional protein within the protein complex consisting of G A R TFase, AICAR TFase, Ser HMase, and trifunctional protein. Resolution of the trifunctional protein from the remaining enzymes in the complex causes a loss of G A R TFase activity which is regained upon recombination. The minimum stoichiometry for G A R TFase reactivation is 3: 1 G A R TFase-trifunctional protein. Determination by ultracentrifugation of the sedimentation
coefficient as a function of protein concentration disclosed that the complex is in mobile equilibrium with the individual proteins. However, mixed G A R TFase-trifunctional protein species can be detected by trapping with cleavable bifunctional cross-linking reagents. Additional support for their interaction is found in the kinetic coupling of the trifunctional protein and G A R TFase activities that leads to a fourfold more efficient formation of formylglycinamide ribotide commencing with formate rather than with 5, 10-methenyl-H4folate triglutamate.
In
Scheme I : Affinity Column Ligands
a recent publication we reported the purification of glycinamide ribotide transformylase' via a mild procedure which led to the copurification of three other enzymes: the trifunctional 5,lO-methenyl-, 5,10-methylene-, and 10-formylH4folate synthetase, aminoimidazolecarboxamide ribotide transformylase, and serine transhydroxymethylase, all of which use reduced folate cofactors (Caperelli et al., 1980). The results suggested a specific association between the enzymes such that they might function in vivo as an enzyme complex. In the present paper we present direct proof for the interaction between two of the enzymes.
0
0
CAR-Sepharose
Experimental Procedures
Materials a,P-Glycinamide ribotide (GAR) was prepared by the method of Chettur & Benkovic (1977). 5-Aminoimidazole4-carboxamide-/3-~-ribofuranosyl5'-monophosphate (Ba2+ salt) (AICAR) was prepared by phosphorylation of the corresponding riboside (Yoshikawa et al., 1967; Murray & Atkinson, 1968). AICAR riboside, tetrahydrofolic acid (H4folate), pig heart fumarase, yeast alcohol dehydrogenase, rabbit muscle pyruvate kinase, yeast glucose-6-phosphate dehydrogenase, and bovine pancreatic ribonuclease were purchased from Sigma Chemical Co., St. Louis, MO. Pteroyl(y-~-glutamyl)~-~-glutamic acid was prepared by the method of Baugh et al. (1970) and converted enzymatically to the (-)-~-H,folate(Glu), according to the methods of Blakley (1960) and Mathews & Huennekens (1960). (=t)-~-5,10C'H-H,folate and (+)-r-5,10-CfH-H4folate(Glu), were prepared as described by Rowe (1968). Sodium [14C]formate (59 mCi/mmol) was purchased from Amersham/Searle. G A R Sepharose was synthesized as previously described (Chettur, 1977). The structure is shown in Scheme I. NADP Sepharose was prepared by the technique of Lamed et al. (1973). Sephadex G-200 (fine) and Sepharose 4B were purchased from Pharmacia Fine Chemical, Uppsala, Sweden. Bio-Gel P-300 ( 5 e 1 0 0 mesh) and all electrophoresis chemicals were products of Bio-Rad Laboratories, Richmond, CA. All From the Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802. Received Sepfember 14, 1979; revised manuscript receiced May 19, 1980. This investigation was supported by Grant GM24129. *Recipient of a National Institutes of Health Postdoctoral Fellowship. 8 Recipient of a National Science Foundation Graduate Fellowship.
0006-2960/80/0419-43 13$01.OO/O
AICAR-Sepharose 0
/I
SN(CH215CN=N--NADP
(rlbosei
H
NA DP-Sep harose
other chemicals were of the highest grade commercially available and purified by standard techniques prior to use. AICAR-Sepharose was synthesized via the following diazo coupling technique (Cohen, 1974) (Scheme I). Sepharose 4B (40 g) was activated at 0 "C with 8 g of C N B r a t p H 11 (maintained with 1 M NaOH) according to the procedure of Nishikawa & Bailon (1975). The washed gel cake was suspended in 20 m L of 0.2 M NaHCO, containing 6 m L of 3,3'-diaminodipropylamine,p H 10.0, and shaken for 16 h a t room temperature. The coupled gel was then washed with 500 m L of HzO, 0.2 M acetic acid, HzO, 0.5 M NaOH, H 2 0 , 0.2 M acetic acid, and 2 L of HzO in that order. A portion of the gel (20 g) was shaken with 20 m L of 1 M ethanolamine ~~
' Abbreviations
used: GAR, glycinamide ribotide; FGAR, Nformylglycinamide ribotide; AICAR, 5-aminoimidazole-4-carboxamide1-P-D-ribofuranosyl 5'-monophosphate; H,folate, tetrahydrofolate; H,folate(Glu)3, tetrahydrofolate triglutamate; 2-ME. 2-mercaptoethanol; Me,SO, dimethyl sulfoxide; GAR TFase, GAR transformylase; AICAR TFase, AICAR transformylase; Ser HMase, serine transhydroxymethylase; trifunctional protein, 5,lO-methenyl-, 5, Iomethylene-, and 10-formyl-H4folatesynthetase (combined); NADP, nicotinamide adenine dinucleotide phosphate; HTP, hydroxylapatite cellulose; NaDodSO,, sodium dodecyl sulfate; DTBP, dimethyl 3,3'-dithiobis(propionimidate); BSA, bovine serum albumin.
0 1980 American Chemical Society
4314
B I OC H E MI S T R Y
a t p H 8.3 for 3 h at room temperature. It was then collected by filtration, washed with 2 L of HzO, and dried by suction. The activated gel was suspended in a solution consisting of 24 m L of 0.2 M sodium borate, p H 9.4, and 16 mL of dimethylformamide (DMF) and cooled at 4 "C. To this was added 0.576 g of p-nitrobenzoyl azide in 7 mL of D M F dropwise with stirring at 4 "C. The suspension was stirred for 1.5 h at 4 O C and 7 h at room temperature. The resulting yellowish gel was washed sequentially with 1.5 L of 1 : l DMF-H20 and 2 L of H 2 0 . For reduction of the nitrophenyl group to the corresponding amine, a suspension of the gel was stirred in 30 m L of a solution containing 0.2 M sodium dithionite and 0.5 M N a H C 0 3 . p H 8.5, for 2 h at 40-45 "C, filtered, and washed with 2 L of HzO. For diazotization ice-cold 0.5 M HC1 (30 mL) and 0.1 M N a N 0 2 (30 mL) were added to the moist gel cake, and the suspension was stirred a t 0 "C for 7 min. The gel was collected, dried by suction, washed with 1.5 L of cold H 2 0 , and then resuspended in 30 mL 0.2 M sodium acetate, p H 5.9, at 4 "C. To the suspension was added 0.5 mmol of AICAR (Ba salt), and the suspension was stirred for 1 h at room temperature. The gel was collected by filtration and washed sequentially with I L of H 2 0 . 1 L of a solution of 0.5 M KCI and 50 m M Tris-HC1 (pH 8.0), and 2 L of HzO. The AICAR incorporation was determined by assaying for organic phosphate according to the method of Chen et al. (1956). A typical analysis showed 12 wmol of phosphate/g of gel. The crude enzyme complex was prepared by the technique of Caperelli et al. (1980). Fractions eluted from the second HTP column with an A T P (5-15 mM) gradient were further purified by affinity chromatography.
Methods Enzyme Assays. Assays of GAR TFase, 5,10-C+H-H4folate cyclohydrolase, lO-CHO-H,folate synthetase, and 5,lOCH2-H4folatedehydrogenase were performed according to Caperelli et a]. (1980) with the following exceptions: ( I ) in the synthetase assay, 2 m M ATP, 4 m M MgCI2, 10 m M sodium formate, and 0.4 m M H4folate were used; (2) i n the dehydrogenase assay a 20% excess of formaldehyde over H4folate was employed. Ser HMase was assayed with 0phenylserine by recording the increase in on the production of benzaldehyde (Schirch & Diller, 1971). AICAR TFase was assayed according to the method of Black et al. (1978) Protein concentration was determined by using a U V biuret technique (Zamenhof, 1957). A standard curve was established with four proteins: RNase A. bovine serum albumin, ovalbumin, and y-globulin. All four gave identical standard curves. Affinity Chromatographj,. The partially purified enzyme loaded onto the AICAR-Sepharose affinity column was from an A T P elution of the second H T P column, which had been desalted on Sephadex (3-25 and dialyzed against the equilibrating buffer (Caperelli et al., 1980). All chromatography buffers contained 10 m M 2-ME, 25% glycerol, and 10% M e 2 S 0 , subsequently referred to as "stabilizers". Columns of AICAR-Sepharose (1-20 mL) were prewashed with several column volumes of 75 mM Tris-HC1 and 75 mM K2HP0,, p H 7.5, plus stabilizers and then equilibrated at 4 "C with 7.5 m M Tris-HC1, p H 7.5, plus stabilizers, The enzyme was loaded onto the column in the same buffer, and the column was then washed with 37.5 m M Tris-HC1. pH 7.5. plus stabilizers until the A,,, was constant and low. The purified complex was then eluted with a solution of 75 m M Tris-HC1 and 75 m M K2HP0,, pH 7.5, containing stabilizers. Active fractions were pooled and MezSO was removed by
SMITH ET A L .
dialysis against 7.5 m M K2HP04,p H 7.5, plus 25% glycerol and 10 m M 2-ME. Finally the enzyme was concentrated to 0.5-2 mg/mL (Azso)if necessary on an Amicon YM-10 ultrafiltration membrane and stored in liquid nitrogen. NADP-Sepharose was precycled as with the AICAR-Sepharose, equilibrated with 7.5 m M Tris-HCI, p H 7.5, plus stabilizers, and loaded with enzyme fractions obtained from elution of the AICAR-Sepharose or the H T P column. The column was then washed with more of the equilibrating buffer until the A,,, was constant and low. G A R TFase, AICAR TFase, and Ser HMase were then coeluted with 100 m M Tris-HC1, pH 7.5, plus stabilizers. Finally, pure trifunctional protein was eluted with 75 m M Tris-HC1, and 75 m M K2HPO,, pH 7.5, plus stabilizers, Fractions containing the three activities and fractions containing the trifunctional protein were dialyzed against 7.5 m M K 2 H P 0 4 ,p H 7.5, containing 25% glycerol and 10 m M 2-ME and concentrated to 1-2 mg/mL in the former and 0.5 mg/mL in the latter case. Gel Filtration. A 110 X 1.8 cm Sephadex (3-200 column was equilibrated at 4 "C with 50 m M K2HP0,, p H 7.5, containing stabilizers. The column pressure head was maintained by 15-18 cm of buffer with a flow rate of 3 mL/h. A 3-mL solution of the complex (0.22 GAR TFase unit) purified by AICAR-Sepharose chromatography also containing 0.5 mg each of rabbit muscle pyruvate kinase, pig heart fumarase, yeast glucose-6-phosphate dehydrogenase, yeast alcohol dehydrogenase, and bovine pancreatic RNase A as molecular weight standards was layered on the column and eluted with the equilibrating buffer. One-milliliter fractions were collected. A 1 15 X 1.8 cm Bio-Gel P-300 column was equilibrated at 4 " C with 50 m M K,HP04, p H 7.5, plus stabilizers. The column pressure head was maintained by 15-20 cm of buffer with a flow rate of 7.2 mL/h. The other experimental protocol \vas identical with that described for Sephadex (3-200. Sucrose Density Gradient Ultracentrifugation. This technique was performed by the method of Martin & Ames (1 96 1 ) in 5-20% sucrose gradients (33 mL of total volume). Gradients were poured at room temperature and allowed to equilibrate at 4 "C overnight. Centrifugation was performed on 0.24.75 mL of enzyme at 4 "C in a Beckman Model L5-50 preparative ultracentrifuge equipped with a SW-27 swinging bucket rotor. Centrifugation a t 25 000 rpm was carried out for 48 h. Molecular weights were determined by using yeast alcohol dehydrogenase as an internal standard. Sedimentation VeVelocity. These experiments were accomplished on a Beckman-Spinco Model E analytical ultracentrifuge equipped with UV optics and scanner. Centrifugation at 55 000 rpm was performed at 20 OC in 75 m M K 2 H P 0 4 , 10 m M 2-ME, and 25% glycerol, pH 7.5. The viscosity and density of the solution at 20 "C were 2.5 1 CPand 1.073 g/mL, respectively. Protein concentrations of 0.1-1.5 mg/mL were used in the presence of the substrate GAR (0.25 mM), and concentrations of 0.5 to 2.0 mg/mL were used in the absence of G A R . Enzyme used in these experiments had been purified through AICAR-Sepharose. Sedimentation coefficients were calculated from plots of In r vs. time (Tanford, 1961). Cross-Linking with DTBP. Enzyme at a concentration of 0.5 mg/mL was dialyzed against deaerated, Nz-saturated 50 m M K 2 H P 0 4and 25% glycerol, pH 7.5, under an N 2 atmosphere. The protein mixture was composed of 3: 1 GAR TFase to trifunctional protein obtained by combining appropriate fractions obtained from NADP-Sepharose chromatography. The pH was then adjusted to 8.3 with 1 N KOH, and 0.25 mg of DTBP was added per mL of enzyme solution. The pH
DE NOVO PURINE BIOSYNTHESIS COMPLEX
VOL.
19,
NO.
4315
18, 1980
Table I: Activity Ratios" before and after AICAR-Sepharose ~~~
~
activity ratio
enzyme
1O-formylH,folate synthetase
ratio before ratio after
10.9 15.6
5,lO-methylene- 5 , l O-methenylH,folate H,folate dehydrogenase cyclohydrolase 6.51 11.8
6.5 6.3
AICAR transformylase
Ser HMase
GAR transformylase
0.46 0.34
8.6 12.2
1 1
All activities are relative to GAR transformylase.
was then readjusted to 8.3. After 45 min under an atmosphere of N2, 1.0 mg of solid NaDodS04 was added per mL of reaction mixture, and the enzyme was denatured 2 h in the absence of 2-ME at 37 OC under an atmosphere of N2. Cross-links were cleaved after the first dimension of electrophoresis with 10% 2-ME in NaDodSO, gel buffer. NaDodS04 Gel Electrophoresis. One-dimensional electrophoresis employed 5 or 7.5% polyacrylamide gels that were run at 8 mA/gel according to Weber et al. (1972). Twodimensional NaDodSO, gel electrophoresis on 5 or 7.5% gel was carried out as described by Coggins et al. (1976). The slab gel apparatus was constructed in our laboratory according to a modification of the apparatus of Driedger & Blumberg (1978). Coomassie blue staining was performed by the method of Weber et al. (1972). Silver staining of slab gels was performed by the method of Merril et al. (1979). Recombination Experiments. Solutions of the trifunctional protein ranging in concentration from 0.02 to 1.0 p M were prepared in media buffered with 7.5 mM K2HP04,25% glycerol, 10 m M 2-ME, and 0.2 mg/mL BSA (pH 7.5). These diluted solutions did not lose enzymic activity overnight. The G A R TFase fraction obtained from NADP-Sepharose elution (92% G A R TFase and 7% trifunctional protein based on densitometer scans of NaDodSO, gels) was diluted to 0.65 and 0.029 p M in G A R TFase and trifunctional protein, respectively, with the above buffer media immediately before use. Recombination was commenced by adding 6.75 pL from solutions containing the trifunctional protein to the GAR TFase assay buffer. Then 6, 9, 12, or 15 pL of 0.65 pM GAR TFase was added, and the resulting solution incubated at 37 OC for 10 min before introducing 5,10-methenyl-H4folate. The final assay volume was 0.675 mL. The GAR TFase reactivation was determined by assaying for its activity. Calculations for the complex dissociation constant ( K d )were based on formal total protein concentrations in the assay medium. A sample of the purified trifunctional protein (50 pL, 0.528 mg of protein/mL) was diluted to 100 p L with 7.5 m M potassium phosphate, 25% glycerol, and 10 m M 2-ME, pH 8.0. The solution was brought to pH 8.3 by addition of 0.1 M KOH and was made 50 mM in iodoacetamide by the addition of the solid. After the solution was mixed, the reaction was allowed to proceed for 45 min at room temperature. The entire solution was dialyzed against 100 mL of 7.5 mM potassium phosphate, 25% glycerol, and 10 m M 2-ME, pH 8.0 followed by dialysis against 100 mL of 7.5 m M potassium phosphate and 25% glycerol, pH 7.5. The recovered protein solution (50 1 L ) was used in recombination experiments with the G A R TFase fraction following the above procedure. Dilution Experiments. The complex eluted from AICARSepharose (2.9 p M G A R TFase and 2.0 pM trifunctional protein based on densitometer scans of NaDodS04 gels) was dissolved in the G A R TFase assay medium to give concentrations of GAR TFase and trifunctional protein ranging from 12.2 to 0.65 n M and 8.4 to 0.45 nM, respectively, and preincubated for 30 min-3 h before 5,lO-methenyl-H,folate was added to initiate the assay.
0
0.2
0 0
0.5 1.0 Relative Mobility
FIGURE 1: NaDodSOP gel electrophoresis of Tris-phosphate elution
from AICAR-Sepharose. Electrophoresis was performed according to the method of Weber et al (1972) on 7.5% polyacrylamide gels. Protein bands are assigned as follows: I, trifunctional protein; 11, A I C A R TFase; 111, G A R TFase, IV, Ser HMase.
Coupled Kinetic Experiments. Experiments designed to demonstrate coupling of the complex activities were performed under N 2 using a reaction mixture containing 0.1 M NH4C1, 2 m M sodium formate, 0.5 m M (a+P)GAR, 4 m M MgC12, 14 p M H4folate(Glu),, and 2 m M ATP in 50 m M maleate, pH 6.8, at 25 "C. In each case the reaction was initiated with 8 pL of enzyme complex: by activities (8.94 X lo-* unit/mL GAR TFase: 1.41 units/mL synthetase) and by densitometer scans of NaDodSO, gels (1.51 pM GAR TFase and 0.910 pM trifunctional protein). The production of 5,l 0-methenyl-H4folate(Glu)3 was followed by AA355. The total of 5,lOmethenyl-H4folate(G1u)3 plus 1O-f0rmyl-H~folate(G1u)~ produced was determined by quenching 50-pL aliquots, removed at various times, in 150 pL of 6 N HC1 and reading the The amount of FGAR produced was determined radiochemically by using 2 m M sodium [14C]formate. In this assay 150-pL portions of the reaction mixture were removed at 0, 15, 30, and 60 min and quenched in 150 p L of 0.093 N NaOH. Each quenched sample was applied to a 0.55 X 9 cm column of QAE-Sephadex which had been equilibrated previously with 0.01 M NaHCO,, pH 9.9. [14C]FGAR was eluted with 0.02 M NaHCO,, pH 9.9. Radioactivity measurements were done with a Beckman LS-8100 liquid scintillation counter using Aquasol-2 (New England Nuclear). Results
Coelution from an AICAR Affinity Column. As has been previously shown, G A R TFase, AICAR TFase, 5,lOmethenyl-H,folate cyclohydrolase, IO-formyl-H4folate synthetase, 5,lO-methylene-H,folate dehydrogenase, and Ser HMase activities copurify on GAR-Sepharose (Caperelli et al., 1979, 1980). The specific association of the various activities suggested by this result is supported by a similar copurification of the enzymes on a column specific for AICAR TFase. However, since elution has been effected by a nonspecific ligand, phosphate, this finding does not constitute strong supporting evidence for the complex. The activities of the proteins obtained by this latter procedure are listed in Table I; a typical protein gel pattern resulting from elution from AICAR-Sepharose is shown in Figure 1. Both the AICARand GAR-Sepharose columns result in purification with
4316
S M I T H ET A L .
B I OC H EM I STR Y I
Table 11: Specific Activities of the Trifunctional Protein
sp act. (units/mg)
x-fold purifn
enzyme synthetase dehydrogenase cyclohydrolase
10.3 12.1 30.6
74 1 681 936
/-
8.0-
from crude
1
I
A5800 0.2
00
0.5
I
5.0
I
15.0
CTrifunctionol Protein], nM.
Relative Mobility
m
FIGURE 3: Activation of GAR TFase by the trifunctional protein. and 14.2 GAR TFase a t concentrations of 5.7 (0),8.5 (A),11.4 (O), nM (V) were recombined in the GAR TFase assay with varying trifunctional protein concentrations. The enzymes were preincubated together at 37 “C for 10 min before 5,10-methenyI-H4folatewas added to start the assay.
o,2tII 0
I
10.0
0.5 Relative Mobility
Upper: NaDodSO, gel electrophoresis of Tris elution of NADP-Sepharose, 92% GAR TFase (relative mobility = 0.35), and 7% trifunctional protein (relative mobility = 0.22). Lower: NaDodS04 gel electrophoresis of Tris-phosphate elution of NADP-Sepharose, 100% trifunctional protein. Electrophoresis was performed according to the method of Weber et al. ( 1 972) on 7.5% polyacrylamide gels. FIGURE2 :
comparable activity ratios for the same four proteins which have been assigned as I, trifunctional protein, 11, AICAR TFase, 111, GAR TFase, and IV, Ser HMase (Caperelli et al., 1980). Since the AICAR-Sepharose capacity is typically greater than that of the GAR-Sepharose (the former binds - 2 mg of protein/mL of resin while the latter binds -0.03 mg/mL), it becomes the preferred method for obtaining these activities. Chromatography on NADP-Sep h arose. Affinity chromatography on NADP-Sepharose of the complex obtained after phosphate (75 mM) elution of the proteins from the AICAR affinity column results in purification to homogeneity of the trifunctional protein and recovery in a subsequent fraction of a protein solution -92% G A R TFase and 7% trifunctional protein with minor amounts of AICAR TFase and Ser HMase. This composition was estimated from densitometer scans of NaDodSO, analytical gels assuming equivalent staining of the various proteins. However, the purification step leads to 80% loss of the G A R TFase activity. The results are summarized in Table I1 and shown in Figure 2. Recombination Experiments. Since removal of the trifunctional protein from G A R TFase on NADP-Sepharose leads to inactivation of the G A R TFase, it was reasoned that readdition of the former may lead to reactivation. The result of titrations of four differing concentrations of G A R TFase with the trifunctional protein are presented in Figure 3, showing a large activation (up to 10-fold) of G A R TFase in all four instances upon saturation by the trifunctional protein. A cursory study of the time dependence for reactivation showed
-
the latter to be complete within the dead time of the assay, I O min. Bovine serum albumin, lactate dehydrogenase, catalase, or pyruvate kinase at concentrations of 0.1 mg/mL in the assay do not replace or inhibit the trifunctional protein in the reactivation process. The latter may be viewed as an equilibrium of unknown stoichiometry governed by the dissociation constant (Kd) according to the equation iV1(GAR TFase) +
2
N2(trifunctional protein) (GAR TFas e) N,(t r i fu nct ional protein ) N 2 where Kd =
(GAR TFase)”l(trifunctional protein)”’ [(GAR TFase),y,(trifunctional protein),,]
(1)
N = number of equivalents. Moreover, the initial velocity for GAR formation is presumed to be linearly proportional to the associated proteins, i.e. v = k[(GAR TFase),v,(trifunctional p r ~ t e i n ) , ~ , ] (2) Employing an interactive computer procedure, values of Kd were varied for assumed stoichiometries 1:1, 2:1, and 3:l (N,-N,) for GAR TFase-trifunctional protein to calculate the (GAR TFase),,(trifunctiOnal protein)N, concentration to give the best least-squares fit to the linear equation, eq 2 . A minimum stoichiometry of 3: 1 G A R TFase-trifunctional protein with an average Kd = 250 f 150 nM3 and k = 261 min-I resulted in the “best fit” with a correlation coefficient for all 32 data points of 0.985 (Figure 4). Since the G A R TFase fraction was not obtained completely free of the trifunctional protein, it was not possible to determine directly whether the former is totally inactive in the absence of trifunctional protein; however, the intercept value for Figure 4 is zero within experimental error. If a true equilibrium is being monitored, the same equilibrium constant should be measurable from either direction of the reaction, i.e., the concentration of the active complex [(GAR TFase)3(trifunctiona1protein)] should vary according to eq 1 when the purified complex (after AICAR-Sepharose chromatography) is diluted through K d . Further, the trans-
V O L . 19,
DE NOVO PURINE BIOSYNTHESIS COMPLEX
NO.
18,
1980
4317
C (GAR TFase& Trifunction I FIGURE 4: Plot of the experimental GAR TFase activity vs. calculated ( G A R TFase),(trifunctionaI protein) conccntration for Kd = 250 nM' and k = 261 min-' according to cq 2.
I
I
5.0 [GAR TFose 1
I
I
10.0
FIGURE 5: Effect of dilution upon GAR TFase activity. Plot of activity vs. GAR TFase concentration ( 0 )and vs. (GAR TFase),(trifunctionaI protein) concentration when Kd = 10 ( 0 ) .I50 (A),and 500 (0)nM'.
The activity was determined after the enzyme complex from the AICAR-Sepharose column had been incubated 30 min under assay conditions in the absence of 5,1O-methenyl-H,folate. The assay was started by adding the cofactor in a small volume of buffer. (Incubation of the enzyme for up to 3 h under the assay conditions in the absence of cofactor prior to assay gave identical results.) formylase activity of the (GAR TFase),(trifunctionaI protein) complex should vary linearly with its calculated concentration but deviate from linearity if the value of Kd and hence the concentration of the complex are in error. The linear iterative computer program used above was modified to accept the dilution data, and Kd was varied from 1.0 to 500 nM' in increments of 10 nM3. A best fit for G A R TFase activity as a function of [(GAR TFase),(trifunctionaI protein)] was found at Kd N 150 nM' in excellent agreement with the above value. This is shown in Figure 5. Cross-Linking. Two-dimensional NaDodSO, gel electrophoresis of proteins linked with cleavable cross-linkers has been the method of choice for the investigations of nearest-neighbor interactions in a number of protein complexes (Coggins et al.,
FIGURE 6: Two-dimensional NaDodSO, gel electrophoresis Of cross-linked complex. The complex was cross-linked as described under Experimental Procedures. The first dimension was run in an I I-cm tube gel (5% polyacrylamide), and the cross-links then were cleaved with 2-ME. The second dimension was run by placing the tube gel on top of a n 1 I x 14 X 0.2 cm slab gel of 7.5% polyacrylamide.
1976; Baird & Hammes, 1976; Smith et al., 1978; Sun et al., 1974). For our study, the thiol cleavable reagent DTBP was chosen for its length (-10 A), water solubility, and imido ester reactivity (Peters & Richards, 1977). Two-dimensional NaDodSO, gel electrophoresis was performed on the cross-linked enzyme, with the result shown in Figure 6 . In addition to the monomer spots for the trifunctional protein, AICAR TFase. and G A R TFase, four major and several minor cross-link (off-diagonal) spots are apparent. The major spots correspond to (A) G A R TFase dimer, (B) AICAR TFase dimer, ( C ) trifunctional protein dimer, and ( C plus D) the heterologous cross-link of 2 G A R TFase units with one trifunctional protein unit. As further evidence, the minor spots for the cross-link of 1 GAR TFase unit (E) with 1 trifunctional protein unit (F) and that for 2 G A R TFase units (G) with 2
4318
B I OC H EM I ST R Y
S M I T H ET A L .
Table 111: Ultracentrifugation Conditions Which Resolve the Comulex Comuonents
buffer constituents, pH 7.5 37.5 niM Tris-HC1, 30% Me,SO, 5% glycerol, 10 mM 2-MEa 37.5 inM Tris-HC1, 30% Me,SO, 5% glycerol, 10 mM 2-1111,' 50 m M Tris-HCI, 10 mhl 2-1111: 10 111M K,HPO,. 10 mbl 2-MI: 50 m M Tris-HC1, 10% Me,SO. 10 inhi 2 4 1 : 10 mhl K,HPO,, 10 mM 2-MIlb
initial [protein] (mg/mL)
mL of protein
2.6
0.50
0.86
0.50
0.86 2.5 0.86
0.50 0.75 0.20
25.0
-
c .synthetase A-A
0.50
In thc presence of glycerol, the elements of the complex did not sediment very far through the sucrose gradient; however, the activity peaks were alu.ays resolved. Ultracentrifugation was performed in 5-2074 sucrose gradients containing the above constituents. This experiment was performed on crude enzyme after protamine sulfate precipitation of nucleic acids to determine if a n y steps in the purification lead to decreased association. No increased asociation \vas seen. Table IV:
cyclohydrolose dehydrogenose
Molecular \\'eight\ for Complex ISnzymes
enzyme
subunit .M,
overall Mr
no. of subunits
trifunctional AICAR transformylase GAR transformylase Ser HMase
90 000-97 000 66000-69000
190 000-200 000 130000-140000
2 2
-
-
.04 .03.02-
-
.05- AICAR Transformylase
+L-&LLJ .oI
2
6
IO
14
18
22
26
30
34
Tube Number
7: Sucrose density ultracentrifugation of complex enzymes. Protein from AICAR-Sepharose was layered onto the 5-20% sucrose gradient. Ultracentrifugation was carried out in 10 rnM K 2 H P 0 , and 10 rnM 2-ME, p H 7.5, at 4 OC for 48 h at 25000 rprn. Upon completion, the tube was fractionated into 1 -rnL fractions. FIGURE
55000-57000
110000-120000
2
47000-50000
190000-215000
4
trifunctional protein units (H) are also apparent. Spots D and G are degenerate with a spot for G A R TFase tetramer and hexamer, respectively; however, cross-linking of the protein fraction containing 92% GAR TFase did not show these spots. Thus, they arise from heterologous interactions with the trifunctional protein. The cross-linking reactions shown here were quenched with NaDodSO, rather than ammonium ion. Quenching with the latter caused the reaction solution to become turbid, and only very high molecular weight and unresolved protein bands were observed on 5% NaDodSO, gels. It is possible that the ammonium causes some association change to occur. At the protein concentration used, 0.5 mg/mL, control proteins do not cross-link. Molecular Weights. It previously had been reported that the protein elements of this complex are resolved by sucrose density ultracentrifugation in 50 m M Tris and 10 m M 2-ME (Caperelli et al., 1980). Since we have noted that glycerol, phosphate, and Me,SO stabilize the enzyme activities, ultracentrifugation experiments were performed in their presence. A typical sedimentation pattern is shown in Figure 7 . Dissociation was observed under all the conditions reported in Table 111. The subunit composition and molecular weights assigned on the basis of this study and N a D o d S 0 4 gel electrophoresis are listed in Table IV. Gel permeation chromatography on either Sephadex G-200 or Bio-Gel P-300 in the presence of stabilizers likewise resulted in partial resolution of the complex into its respective activities but did not provide reliable estimates of their molecular weights. Sedimentation velocity and equilibrium experiments can give the overall molecular weight of a protein complex if the individual proteins do not dissociate in the centrifugal field. We have shown that dissociation does indeed occur in a sucrose gradient a t 25 000 rpm. Analytical sedimentation velocity
Table V: Effect of Complex Concentration upon Sedimentation Coefficient in the Presence of GAR
0.10 0.30 0.60 0.60 1.43
6.94 7.06 7.16 7.15 7.25
experiments were performed in an attempt to determine if association would be maintained lacking a sucrose gradient. In the absence of the substrate GAR, no distinct UV boundary was formed a t 0.5-2.0 m g / m L protein. Thus, lack of association is again suggested. However, in the presence of 0.25 m M (a+P)GAR, an apparent single boundary does form, suggesting that G A R causes association. The effect of protein concentration on the sedimentation coefficient was investigated in order to determine s20,wextrapolated to zero protein concentration. A plot of szo,wvs. protein concentration gives a negative slope for nondissociating macromolecules, and the s20,w extrapolated to zero protein concentration is the true sedimentation coefficient (Fujita, 1962). However, a rapidly reversing associating system can lead to a plot with a positive slope (Gilbert & Gilbert, 1973; Schwert, 1949; Schachman, 1959). The sedimentation coefficients for the present system are shown in Table V . It is apparent that s20,wdoes increase with increasing protein concentration, implicating a rapidly associating and dissociating complex. Coupling of Activities. Since the above findings were in accord with a physical interaction between the G A R TFase and the trifunctional protein, a kinetic test was undertaken of the ability of the complex to synthesize FGAR commencing
V O L . 19,
DE NOVO PURINE BIOSYNTHESIS COMPLEX
Scheme I1 HOOC-Na+
-I- A T P i- H4folate(Glu)3
e -
lO-CHO-H4folate(Glu)3 5.10-C+H-H4folate(Glu)3
-I- GAR
" C 80
4319
1 8 , 1980
5.10-C+H-H~folate(Glu)~
FGAR iHqfolate(Glu)3
1
160-
2
NO.
i b A A ~ A AA A A
A
A
'0
15
20
5
IO
A
A
25
30
A
35
40
45
50
TIME ( m i d
/
TIME (rnin)
FIGURE8: Rate of production of FGAR by the complex under kinetic
coupling conditions using 14 p M H4folate triglutamate and 2 m M sodium [I4C]formate (-) and from the G A R TFase assay using 2.5 pM 5,lO-methenyl-H4folate triglutamate (- - -).
with formate, ATP, and H,f~late(Glu)~, thus demanding the participation of activities associated with the trifunctional protein (see Scheme 11). In Figure 8 is plotted the linear rate of FGAR synthesis upon initiating its formation from formate relative to .5,1O-C+H-H,folate(Gl~)~. The former pathway is 4 times more efficient than that beginning with exogenous 5,10-CfH-H4-folate(Glu),, the latter concentration set at 2.5 p M based on the steady-state level attained by this cofactor species during the course of F G A R synthesis (Figure 9). Neither A T P nor formate activates G A R TFase under these conditions. The ratio of 10-formyl-H4folate(Glu)3to 5,lOC'H-folate in solution is 4.1 :1, approaching the 7: 1 calculated for p H 6.8 from the reported equilibrium constant (Poe & Benkovic, 1980). In order to test further the kinetic coupling as well as the activation upon recombination involving the trifunctional protein and G A R TFase, an iodoacetamide-modified trifunctional protein was obtained which had no measurable synthetase activity and only very low levels of cyclohydrolase and dehydrogenase activity. The effect of the iodoacetamide modification on G A R TFase activity in the recombination experiment is summarized in Table VI. Neither the modified nor unmodified trifunctional protein showed any significant G A R TFase activity when assayed separately. At concentrations twofold greater than that saturating with the unmodified protein, inactive trifunctional protein does not activate G A R TFase. A control experiment in which G A R TFase was assayed in the presence of both modified and native trifunctional protein shows that reactive ligands possibly carried over from the modification reaction are not responsible for the lack of G A R TFase activation when assayed with the modified trifunctional protein. Discussion The copurification of G A R TFase, AICAR TFase, Ser HMase, and the trifunctional protein possessing 5,lOmethenyl-H,folate cyclohydrolase, 5,lO-methylene-H4folate dehydrogenase, and 10-formyl-H,folate synthetase activities
FIGURE 9 : Concentrations of H,folate cofactors produced by the complex under kinetic coupling conditions from 14 p M H,folate triglutamate and 2 m M sodium formate: (0) 10-formyl- plus S,lO-methenyl-H,folate triglutamate measured by after acidification; (A)5,10-methenyl-H4folate triglutamate measured by AA355; (0)1 O-formyl-H4folate triglutamate determined by the difference between the total of 10-formyl- plus 5,10-methenyl-H4folate triglutamate and 5,10-methenyl-H4folate triglutarnate alone.
Table VI: Recombination Experiment with Modified Trifunctional Protein [GAR TFase] (nM)
iodoacetamidemodified [ trifunctional protein] (nM)
native [ trifunctional protein] (nhl)
re1 act.
12 12 12 12
0 26 0 26
0 0 12 12
0.22 0.27 1.o 1.1
has been previously reported (Caperelli et al., 1980). The suggestion of an association between the enzymes is further supported by the results presented here. The utility of such an enzyme complex is apparent, since under complexing conditions the unstable H,folate cofactors may not dissociate into solution prior to their use by the G A R TFase or AICAR TFase. Furthermore, the complex may furnish a convenient means for regulation of de novo purine biosynthesis. Direct evidence for the interaction between the enzymes of this complex can be seen in the recombination, dilution, cross-linking, sedimentation velocity, and kinetic coupling experiments. The recombination and dilution experiments showed that the trifunctional protein affects a necessary and specific activation of GAR TFase. On the basis of the ordinate intercept of zero in the plot of GAR TFase activity vs. complex concentration, G A R TFase would be inactive in the absence of the trifunctional protein. This activation must be due to a direct physical interaction between the proteins. The cofactor for G A R TFase, 5,lOmethenyl-H,folate is unstable under the assay conditions of p H 7.0, 50 m M potassium maleate, hydrolyzing with a t I l z of 30 min (P. A. Benkovic, unpublished data) to 10-formylH,folate which is not used by the enzyme (Hartman & Buchanan, 1959). This hydrolysis is also catalyzed by the 5 , IO-methenyl-H4folate cyclohydrolase activity of the trifunctional protein, and the specific activity of this enzyme is typically 10 times that of the G A R TFase. Thus, any indirect affect of the presence of the trifunctional protein should be to inhibit the G A R TFase reaction by removing cofactor. Obviously this is not the observed effect, since the enzyme indeed activates G A R TFase. The equilibrium for this interaction is rapidly established as demonstrated by the reactivation and dilution experiments. It was found that the G A R TFase is completely reactivated
4320
B I oc H E M I sT R Y
SMITH ET A L .
Scilellle I11 (GAR TFase)3(trifunctional protein)
t rifunctional
+
GAR TFase
/ e (GAR TFase)2(trifunctlonaI protein)
protein (GAR TFase) (trifunctional protein)
by a 10-min preincubation (the experimental dead time) with the trifunctional protein prior to commencing the assay with 5,10-methenyl-H4folate. Similarly, the dilution experiments indicated that the equilibrium in the direction of dissociation is established within 30 min. Furthermore, the specificity of this recombination was demonstrated clearly through the lack of activation with several nonspecific proteins and inactivated trifunctional protein. That the latter does not affect activation additionally may indicate that processing of the cofactor species at sites on the trifunctional protein may be important. This will be the subject of a future publication. The reversible cross-linker DTBP introduced chemical links between GAR TFase and the trifunctional protein, indicating direct interaction between the two enzymes. Thus, the cross-linking experiments are in accord with the reactivation of G A R TFase by the trifunctional protein which indicated that the two enzymes should be nearest neighbors. The major heterologous cross-link was that of two GAR TFase subunits with one trifunctional protein subunit. The amount of protein in the spot corresponding to this cross-link is comparable to that in the G A R TFase dimer spot. suggesting that the enzymes interact strongly at 0.5 rng/mL. This is an anticipated result based on the recombination dissociation constant of 250 m M 3 (or 6.3 m M for the 1 : 1 interaction) so that association must be >90% a t 0.5 mg/mL. The apparent lack of cross-links between AICAR TFase and either G A R TFase or the trifunctional protein should not as yet be interpreted to rule out a n association between this transformylase and other proteins in the complex. Specifically, the AICAR TFase mole ratio to GAR TFase was -0.05 in this experiment; thus, cross-linking involving AICAR TFase might not be seen. The cited cross-linking reactions were performed at pH 8.2, and it is not clear how this alkaline pH affects the protein association. Thus, at lower pH more extensive cross-linking may have been observed. although such experiments are precluded owing to more rapid imido ester hydrolysis (Hunter & Ludfiig. 1962). Although the cross-linking and recombination experiments show that GAR TFase and the trifunctional protein do indeed interact, their association must be rapidly reversible. This is supported by the sedimentation velocity, sucrose density sedimentation, and gel filtration experiments. In particular, velocity sedimentation of a system featuring rapidly reversible association between macromolecules leads to- an increasing sedimentation coefficient with increasing protein concentration in dilute solution (Gilbert & Gilbert, 1973; Schachman. 1959). This characteristic presumably results from the larger population of aggregated species at higher protein concentrations due to mass action, and thus, measurements on this type of system which sediments as the weight average yield a higher apparent sedimentation coefficient with increased protein concentration (Gilbert, 1963; Schachman, 1959). This is the effect seen with the present case in accord with rapid association-dissociation. Thus. an overall molecular weight analysis cannot readily be performed on the data. Similar results are also apparent i n both the sucrose density ultracentrifugation and gel filtration experiments. Here anomalous elution or incomplete resolution indicates the lack of a stable association.
The available data can be related to the outline in Scheme 111. This model is supported by the observations that (1) GAR TFase is activated by 1, 2/3, or 1/3 trifunctional proteins, (2) cross-linking experiments reveal the presence of 1 : 1, 2: 1, and 2:2 G A R TFase monomer(s)/trifunctional protein monomer(s), and (3) sedimentation analysis indicates rapid reversibility. It is important to realize that for catalytic activity only one GAR TFase need be associated with the trifunctional protein at a given time. This ratio may be sufficient to produce a more efficient FGAR synthesis as noted in experiments commencing with formate, ATP, and H , f o l a t e ( G l ~ )or ~ to result in the inability of iodoacetamide-modified trifunctional protein to activate G A R TFase. Other examples of enzymes which only complex under specific conditions are known. The enzyme complex involved in the early steps of de novo pyrimidine biosynthesis shows tight association only in the presence of MezSO and glycerol (Mori & Tatibana, 1978; Coleman et al., 1977; Mori et a]., 1975). I n the absence of MezSO and glycerol the large complex of -870 000 dissociates into smaller units. Similarly, the interaction of aspartate aminotransferase (AAT) and malate dehydrogenase ( M D H ) can only be demonstrated under specific conditions. Physical evidence for the complex is obtained from countercurrent distribution in 6.4% dextran, 6.6% trimethylaminopoly(ethylene glycol), or carboxymethylpoly(ethylene glycol) (Backman & Johansson, 1976). In the absence of these constituents, kinetic evidence for the interaction is observed but physical evidence is not. Indeed, chromatography of MDH on a Sephadex 6 2 0 0 column equilibrated with AAT caused no anomalous elution of the MDH (Bryce et al., 1976). Halper & Srere (1977) have found that M D H and citrate synthase also show specific association in the presence of poly(ethy1ene glycol) whereas no apparent interaction exists i n its absence. Of particular interest is the recent discovery of a tetrahydrofolate synthesizing multienzyme complex in Escherichia coli (Toth-Martinez et al., 1975). It would be important to determine if such a complex exists in eukaryotes and if this system interacts Rith the purine complex. Acknoa ledgments We thank Brian Cunningham for writing the original version of the computer program. References Backman, L., & Johansson, G . (1976) FEBS Lett. 65, 39. Baird. B. A,, & Hammes, G . G . (1976) J . Biol. Chem. 251, 6953. Baugh, C . M., Stevens, J. C., & Krumdieck, C. ( 1 970) Biochim. Biophys. Acta 212, 116. Black. S. L., Black, M. J., & Mangum, J. H. (1978) Anal. Biochenz. 90, 397. Blakely, R. L. (1960) Nature (London) 188, 231. Bryce, C. F. A,, William. D. C., John, R . A., & Fasella, P. (1 976) Biochem. J . 153, 57 1. Caperelli, C . A,, Chettur, G., Lin-Kosley, L., & Benkovic, S. J . ( 1 979) in Chemistry and Biology of Pteridines (Kisliuk, R. L., & Brown, G. M., Eds.) p 371, Elsevier/North Holland, Amsterdam. Caperelli, C. A., Benkovic, P. A., Chettur, G., & Benkovic, S. J . (1980) J . Biol. Chem. 255, 1885. Chen, P. S., Jr., Toribara, T. Y., & Warner. H. (1956) Anal. Chenz. 28. 1756. Chettur, G . (1977) Ph.D. Thesis, The Pennsylvania State Universit). Chettur, G., & Benkovic, S. J. (1977) Carbohydr. Res. 56, 75.
Biochemistry 1980. 19, 4321-4327 Coggins, J. R., Hooper, E. A., & Perham, R. N. (1976) Biochemistry 15, 2527. Cohen, L. A. (1974) Methods Enzymol. 34, 102. Coleman, P. F.,Suttle, D. P., & Stark, G. R. (1977)J . Biol. Chem. 252, 6379. Driedger, P. E.,& Blumberg, P. M. (1 978) Anal. Biochem. 87, 177. Fujita, H . (1962) Mathematical Theory of Sedimentation Analysis, Academic Press, New York. Gilbert, G. A. (1 963) Ultracentrifugal Analysis in Theory & Experiment (Williams, J. W., Ed.) pp 73,Academic Press, New York. Gilbert, L. M., & Gilbert, G. A. (1973)Methods Enzymol. 27, 273. Halper, L. A., & Srere, P. A. (1 977)Arch. Biochem. Biophys. 184, 529. Hartman, S. C., & Buchanan, J. M. (1959)J . Biol. Chem. 234, 1812. Hunter, M. J., & Ludwig, M. L. (1962)J . A m . Chem. SOC. 84, 3491. Lamed, R., Levin, Y., & Wilcheck, M. (1973) Biochim. Biophys. Acta 304, 23 1. Martin, R.G., & Ames, B. N. (1961) J . Biol. Chem. 236,
1372. Mathews, C. K., & Huennekens, F. M. (1960)J. Biol. Chem. 235, 3304. Merril, C. R., Switzer, R. C., & Van Keuren, M . L. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 4335.
4321
Mori, M., & Tatibana, M. (1978)Eur. J . Biochem. 86, 38 1. Mori, M., Ishida, H., & Tatibana, M. (1975)Biochemistry 14, 2622. Murray, A. W., & Atkinson, M. R. (1 968) Biochemistry 7 ,
4023. Nishikawa, A. H., & Bailon, P. (1975) Anal. Biochem. 64,
268. Peters, K., & Richards, F. M. (1977)Annu. Rev. Biochem. 46, 523. Poe, M., & Benkovic, S.J. (1980) Biochemistry (in press). Rowe, P. B. (1968) Anal. Biochem. 22, 166. Schachman, H. K.( 1959) Ultracentrifugation in Biochemistry, Academic Press, New York. Schirch, L., & Diller, A. (1971) J . Biol. Chem. 246, 3961. Schwert, G. W. (1949) J . Biol. Chem. 179, 655. Smith, R. J., Capaldi, R. A., Muchmore, D., & Dahlquist, F. (1978)Biochemistry 17, 3719. Sun, T. T., Bollen, A., Kahan, L., & Traut, R. R. (1974) Biochemistry 13, 2334. Tanford, C. (1 96 1) Physical Chemistry of Macromolecules, Wiley, New York. Toth-Martinez, B. L., Papp, S., Dinya, Z., & Hernadi, F. J. (1 975) BioSystems 7, 172. Weber, K., Pringle, J. R., & Osborn, M. (1972) Methods Enzymol. 26, 3. Yoshikawa, M., Kato, T., & Takenishi, T. (1967)Tetrahedron Lett., 5065. Zamenhof, S. (1957) Methods Enzymol. 3, 696.
Methotrexate-Resistant Chinese Hamster Ovary Cells Contain a Dihydrofolate Reductase with an Altered Affinity for Methotrexatet Wayne F. Flintoff* and Karim Essani
ABSTRACT:
Previous reports [Flintoff, W. F., Davidson, S. V., & Siminovitch, L. (1976)Somatic Cell Genet. 2, 245-261; Gupta, R. S.,Flintoff, W. F., & Siminovitch, L. (1977)Can. J. Biochem. 55,445-4521 described a series of Chinese hamster ovary cells that were resistant to the cytotoxic action of methotrexate and contained a dihydrofolate reductase that was less sensitive to inhibition by the drug than wild-type enzyme. In this study, binding of labeled methotrexate to the reductase-NADPH complex and separation of free and bound drug by filtration through Sephadex G-25 have been used to demonstrate that clonal isolates of these resistant cells contain a dihydrofolate reductase varying between 2.5-and 6-fold lower
in affinity for the drug than the wild-type enzyme. The apparent dissociation constant for the wild-type enzyme is 0.5 X 1 OW9M. Using two-dimensional polyacrylamide gel electrophoresis, 1 1 independently selected resistant isolates have been shown to contain a reductase with a similar overall net charge as the wild-type enzyme. Reductase purified from either wild-type or resistant cells contains two components after isoelectric focusing in polyacrylamide gels. The major component represents about 90% of the total protein and has a p l of about 8.0. The minor component representing about loOh of the reductase protein has a p l between 7.2 and 7.6.
E e v i o u s reports from this laboratory have described a series of Chinese hamster ovary cells (CHO)' that have been selected for resistance to the folic acid analogue methotrexate (Mtx) (Flintoff et al., 1976a,b;Gupta et al., 1977). Resistance in class I cells is apparently due to a structural alteration in dihydrofolate reductase, whereas the resistance in class I1 cells involves a defect in the permeability to Mtx. Class I11 cells,
which were derived from class I cells by a second-step selection in an increased concentration of drug, showed increased levels of the enzyme found in class I cells. The conclusion that class I cells are resistant because of a structural alteration in the reductase was based on the increased resistance to Mtx inhibition shown by the reductase
From the Department of Microbiology and Immunology, University of Western Ontario, London, Ontario, Canada, N6A 5C1. Received December 19, 1979. This work was supported by a grant from the Medical Research Council of Canada. W.F.F. is the recipient of a Medical Research Council of Canada Scholarship. K.E. is a Government of Pakistan Overseas Scholar.
Abbreviations used: CHO, Chinese hamster ovary: DFBS, dialyzed fetal bovine serum; Me2S0, dimethyl sulfoxide; EMS, ethyl methanesulfonate; HPRT, hypoxanthine phosphoribosyltransferase; IF, isoelectric focusing; Kd, apparent dissociation constant; Mtx, methotrexate; NADPH, nicotinamide adenine dinucleotide phosphate, reduced form; NG, N-methyl-N'-nitro-N-nitrosoguanidine; PPO,2,5-diphenyloxazole.
0006-2960/80/0419-4321$01 .OO/O 0 1980 American Chemical Society
